Research Of The Unit Process Energy Consumption Model Based On Neural Network

As development of the technology and progress of the social, the energy savingand the environmental protection become more important topic now.The energy isimportant safeguard to the company.The effective use of energy and the reasonablesupply become one of important ways realizing energy conservation.This paper is divided into two parts. In the first part, a set of modem powerenergy monitor system is established.In the second part, artificial neural networkmodel of the unit process energy consumption based on the monitor system.In the first part, establishment of the power energy computer monitor systembased on RS-485 bus is introduced. The instrumentations communicate data utilizingmulti-agreement on the bus.Between the PC and the intelligent instrumentations datacommunicate by means of the bus. On the PC, the KINGVIEW configurationsoftware for the system is utilized.Three level of energy measurement monitormanagement system is constituted. This system has been used in the JMC energymeasurement management system.The system realizes many functions, such asenergy measurement centralization management, automatic data acquisition,real-time data monitor, cumulative inquiry, historical inquiry and print statiscalreport and so on. The system enhances level and intensity of the measurementmanagement, brings great convnience to the production scheduling and measurementmanagement, improves labor productivity effectively and improves the level scale ofthe energy measurement management and contributes energy saving to the company.In the second part, the establishment of artificial neural network model of theunit process energy consumption is introduced. There are many factors in influenceof unit process energy consumption. Because the principal component analysis haslarge ability in choosing main influence factors from relative many factors, principalcomponent analysis is utilized to choose major factors. The unit process hastime-variable, the nonlinear system characteristic. RBF network is a feed-forwardnet of good merit, have nature of functional approximation. RBF network not only has the nature of local approximation and universal optimize, but also has the natureof strong ability in study, nonlinear, generalization ability and far-speeding. RBFneural network is suitable for researching energy consumption model, so the RBFneural network is utilized to established artificial neural network model of the unitprocess energy consumption.The production of the paper follows: firstly, a large energy monitor system iseatablished in JMC limited company.According to the show of data report operatingmore than one year, the purpose of energy saving is achieved and good economicbenefit is obtained.Secondly, the model of unit process energy consumption isestablished in automobile enterprise, it is a useful attempt to solve practical problemson the production using advanced algorithm.